Cardiovascular and Metabolic Risk

O R I G I N A L A R T I C L E

Prevalence of Metabolic Syndrome

Among an Urban Population in Kenya
LYDIA U. KADUKA, MSC1 JOHN K. BORE, MSC4 However, noncommunicable diseases
YERI KOMBE, MD, PHD1 ZIPPORAH N. BUKANIA, MSC
1
(NCDs) contribute over one-half of the
EUCHARIA KENYA, PHD2 MOSES MWANGI, MSC1 top 20 causes of morbidity and mortality
ELIZABETH KURIA, PHD3 (3). Total mortality attributed to NCDs
rose from 31.8% in 2002 to 33% in
2007. This has been attributed to urbani-
OBJECTIVEdDeveloping countries are undergoing an epidemiologic transition accompa- zation that brings with it changes in life-
nied by increasing burden of cardiovascular disease (CVD) linked to urbanization and lifestyle
modications. Metabolic syndrome is a cluster of CVD risk factors whose extent in Kenya
style that adversely affect metabolism.
remains unknown. The aim of this study was to determine the prevalence of metabolic syndrome In 1993, a hospital-based study in
and factors associated with its occurrence among an urban population in Kenya. Nairobi found high prevalence of obesity,
hypercholesterolemia, cigarette smoking,
RESEARCH DESIGN AND METHODSdThis was a household cross-sectional survey and electrocardiogram evidence of left
comprising 539 adults (aged $18 years) living in Nairobi, drawn from 30 clusters across ve ventricular hypertrophy, with hyperten-
socioeconomic classes. Measurements included waist circumference, HDL cholesterol, triacyl- sion as the most common discharge diag-
glycerides (TAGs), fasting glucose, and blood pressure. nosis (4). A study by Christensen et al. (5)
RESULTSdThe prevalence of metabolic syndrome was 34.6% and was higher in women than found the prevalence of overweight
in men (40.2 vs. 29%; P , 0.001). The most frequently observed features were raised blood (BMI $25 kg/m2) at 39.8 vs. 15.8% and
pressure, a higher waist circumference, and low HDL cholesterol (men: 96.2, 80.8, and 80%; obesity (BMI $30 kg/m2) at 15.5 vs. 5.1%
women: 89.8, 97.2, and 96.3%, respectively), whereas raised fasting glucose and TAGs were among urban versus rural Kenyan popula-
observed less frequently (men: 26.9 and 63.3%; women: 26.9 and 30.6%, respectively). The tions, respectively. The same study observed
main factors associated with the presence of metabolic syndrome were increasing age, socioeconomic an overall age-standardized prevalence of
status, and education. diabetes and impaired glucose tolerance of
CONCLUSIONSdMetabolic syndrome is prevalent in this urban population, especially 4.2 and 12.0%, respectively (6). These
among women, but the incidence of individual factors suggests that poor glycemic control is ndings, in addition to the prevailing
not the major contributor. Longitudinal studies are required to establish true causes of metabolic global understanding of CVDs, call for
syndrome in Kenya. The Kenyan government needs to create awareness, develop prevention screening and early detection of metabolic
strategies, and strengthen the health care system to accommodate screening and management of abnormalities to help identify people who
CVDs. are at risk and most likely to benet from
intervention efforts.
Diabetes Care 35:887893, 2012 The term metabolic syndrome refers
to the clustering of a number of cardio-

T
he global prevalence of leading Kenya is a rapidly developing country vascular risk factors (obesity, hyperten-
chronic diseases is increasing, with of sub-Saharan Africa, where the extent of sion, dyslipidemia, and hyperglycemia)
the majority occurring in low- and most cardiovascular diseases (CVDs) and believed to be related to insulin resis-
middle-income countries, and expected to the associated risk factors at population tance. It is estimated that ~2025% of the
rise substantially over the next two decades level remain largely unknown. Chronic worlds adult population have metabolic
(1). Chronic diseases are responsible for diseases have not received much attention syndrome, and they are twice as likely to
50% of the total disease burden, with esti- due to overemphasis on communicable die of and three times as likely to have a
mated age-standardized death rates being diseases, underreporting, missed diagnosis, heart attack or stroke compared with peo-
higher for men and women from low- misdiagnosis, and misclassication of dis- ple without the syndrome. In addition,
income compared to middle-income eases. According to the Ministry of Health people with metabolic syndrome have a
countries (2). Major causes are said to be Annual Status Report 2007, the leading vefold greater risk of developing type 2
increasing rates of hypertension, dyslipi- causes of deaths in Kenya are malaria, diabetes (7). The aim of this study was to
demia, diabetes, obesity, physical inactiv- pneumonia, HIV/AIDS, diarrhea, anemia, determine the prevalence and factors as-
ity, and tobacco use. tuberculosis, meningitis, and heart failure. sociated with the occurrence of metabolic
syndrome among an urban population in
c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c Kenya.
From the 1Centre for Public Health Research, Kenya Medical Research Institute, Nairobi, Kenya; the 2De-
partment of Biochemistry and Biotechnology, Kenyatta University, Nairobi, Kenya; the 3Department of
RESEARCH DESIGN AND
Foods, Nutrition, and Dietetics, Kenyatta University, Nairobi, Kenya; and the 4Directorate of Population METHODSdThe study was carried
and Social Statistics, Kenya National Bureau of Statistics, Nairobi, Kenya. out in Langata constituency of Nairobi
Corresponding author: Lydia U. Kaduka, lydia.kaduka@gmail.com. province, the capital city of Kenya. Langata
Received 17 March 2011 and accepted 3 January 2012. constituency covers an area of 223.4 km2
DOI: 10.2337/dc11-0537
2012 by the American Diabetes Association. Readers may use this article as long as the work is properly and has a density of 1,284 dwellings per
cited, the use is educational and not for prot, and the work is not altered. See http://creativecommons.org/ unit area, a population of 355,188, and a
licenses/by-nc-nd/3.0/ for details. total of 108,477 households. This is in

care.diabetesjournals.org DIABETES CARE, VOLUME 35, APRIL 2012 887

Metabolic syndrome in urban Kenya
comparison with the entire Nairobi prov-
ince, which covers an area of 696 km2
and has a density of 4,509 dwellings per
unit area, with a total of 985,016 house-
holds, and a population of 3,138,369.
The countrys population has grown from
28,686,607 in 1999 to 38,610,097, of
which ~32% live in urban areas (8). Langata
was chosen because of its cosmopolitan na-
ture, and it hosts all the ve socioeconomic
classes (upper, lower upper, middle, lower
middle, and lower), as stratied by the
Kenya National Bureau of Statistics.
This was a cross-sectional study based
on a three-stage, cluster-sampling meth-
odology. The rst stage involved the se-
lection of 30 clusters from the total
enumeration areas in the constituency,
followed by selection of the households,
and nally selection of the respondents.
The study included urban dwellers of
both sexes aged $18 years, with no
known medical history of a debilitating
disease and having resided in Nairobi
for at least 2 years before the study.
Sample size estimation and
sampling
A comprehensive sampling frame was
constructed to represent the target pop-
ulation in which the primary sampling
unit was an enumeration area as dened
during the 1999 Kenya Population and
Housing Census. Enumeration areas are
small nonoverlapping units dened ac-
cording to a specied measure of size and
have maps showing the boundaries and
structures as well as total households and
population by sex. A measure of size is
dened as having an average of 100
households with a lower and upper limit
of 50 and 149 households, respectively. A
total of 30 clusters from all the ve so-
cioeconomic divisions were selected using
the systematic probability-proportional-to-
size sampling method. A quick household
count was carried out in each enumeration
area to verify the number of the entire
households. Enumeration areas that sur-
passed the maximum number of house-
holds were segmented and only one
segment randomly selected. The selected
segments constituted clusters for the
study.
During the sampling process, three
weights were generated from selection
probabilities of clusters, households, and
individuals. The nal individual weight
(taken as inverse of the three selection
probabilities) was applied to the nal data
in order to be representative of the target
population.
888 DIABETES CARE, VOLUME 35, APRIL 2012
Selection of households and
respondents
Following the quick household count, a
household list was generated for every
cluster that included information on
household occupants. The total number
of households per cluster was selected
using the systematic random-sampling
method, followed by the selection of
only one eligible respondent per house-
hold using the Kish Grid technique (9),
which allows selection of one person to
interview from all eligible household res-
idents. Sensitization, recruitment, and as-
sessment of study subjects were done in
three phases. Sensitization at administra-
tive and community levels was done dur-
ing the mapping and counting exercise,
followed by visits to the selected house-
holds and respondents to explain the na-
ture of the study, the requirement for
fasting blood samples, and the intended
day and time of assessments. The third
phase involved a visit to the cluster by
the medical team for blood collection
and assessments.
Assessments
The following assessments were carried
out between August and October 2008.
Socioeconomic assessment. Interviews
were carried out and information on
socioeconomic and demographics col-
lected using a structured questionnaire.
Education level was categorized into ve
levels: none, primary (18 years), second-
ary (914 years), university (.14 years),
and adult education. Occupation was
classied as formal, self-employed, and/
or petty trade. Information on monthly
income also was obtained. Differences in
socioeconomic status were obtained us-
ing the wealth quintiles. The study used
assets ownership and the principal com-
ponents analysis procedure in SPSS to
compute the index. The cut points in
the wealth index at which to form the
quintiles were calculated by obtaining a
weighted frequency distribution of
households and the sampling weight of
the household. This resulted in the ve
categories (i.e., lowest, second, middle,
fourth, and highest wealth quintiles).
Anthropometric assessments. Waist
circumference was determined using
Roche circumference tapes. Subjects were
asked to stand upright in a relaxed manner,
with their feet comfortably apart, their
weight evenly balanced on both feet, and
with their arms hanging by their sides.
Waist circumference was measured at the
point halfway between the lower border of
the ribs and the iliac crest in a horizontal
plane. Measurements were recorded to the
nearest 0.1 cm.
Clinical assessment. Blood pressure was
taken using the automated Omron M6
Comfort (HEM-7000-E). Blood pressure
was measured with subjects in a sitting
position and the arm horizontally sup-
ported. Family history of diabetes and
CVDs were self-reported.
Biochemical assessments. Fasting ve-
nous blood was collected between 7:00
A.M. and 8.00 A.M., and blood glucose was
immediately determined using the Hemo-
cue B-Glucose 201+ analyzer (HemoCue,
Angelholm, Sweden). Blood lipid prole
(triacylglycerides [TAGs], total choles-
terol, and HDL cholesterol) was analyzed
on a Konelab autoanalyzer using the T
Series Triglycerides Kit (cat no. 981301),
the T Series Cholesterol Kit (cat no.
981812), and the T Series HDL Cholesterol
Kit (cat no. 981823) in accordance with the
manufacturers instructions.
Denition of the metabolic
syndrome
Metabolic syndrome was dened accord-
ing to the new 2009 International Di-
abetes Federation consensus statement
criteria, which is the presence of any three
of the ve following risk factors. These are
waist circumference $94 cm for men or
$80 cm for women, TAGs $1.7 mmol/L
or specic treatment for this abnormality,
HDL cholesterol ,1.0 mmol/L for men or
,1.3 mmol/L for women or specic treat-
ment for this abnormality, elevated blood
pressure $130/85 mmHg or treatment of
previously diagnosed hypertension, and el-
evated fasting glucose $5.6 mmol/L or
treatment of previously diagnosed diabetes
(10). The European cutoffs for waist circum-
ference were used because of the unavail-
ability of cutoffs for sub-Saharan Africans.
Ethical considerations
Permission to carry out this study was
sought from the Kenya Medical Research
Institute (KEMRI) National Ethical Re-
view Committee, and informed consent
was obtained from each subject prior to
participation in the study.
Statistical methods
Data were weighted and analyzed using
SPSS version 16, with P , 0.05 consid-
ered statistically signicant. Results are
expressed as means 6 SD or as propor-
tions (%). For categorical variables, the x2
test and Fisher exact probability were
used. Linear associations were calculated
care.diabetesjournals.org

using the Spearman correlation coef-
cient. Binary logistic regression using the
backward conditional method was per-
formed on multiple factors to eliminate
confounding and to examine the effect
of the independent predictors of meta-
bolic syndrome.
RESULTSdA total of 539 (men: 50.5%;
women: 49.5%) participants were as-
sessed for risk factors for metabolic
syndrome. The mean age was 38 6 13
years. Up to 98% of the study partici-
pants were black Africans from the fol-
lowing ethnic groups: Luo (16.6%),
Bohoran (3.7%), Luhya (22.3%), Kalenjin
(1.8%), Kikuyu (11.5%), Kamba (12.7%),
Meru (2%), Kisii (7.4%), Mijikenda (2%),
Nubian (8.6%), Somali (2.6%), Indian
(1.2%), and others (Tanzania, Uganda,
Ethiopia, or Arab) 7.6%. The overall prev-
alence of metabolic syndrome was 34.6%
and was higher in women than in men
(40.2 vs. 29%, respectively; P , 0.001).
Table 1 shows the prevalence of each com-
ponent of the metabolic syndrome by sex
in the study population.
Among those with metabolic syn-
drome, the most frequently observed fea-
tures were raised blood pressure, a higher
waist circumference, and low HDL choles-
terol (men: 96.2, 80.8, and 80%; women:
89.8, 97.2, and 96.3%, respectively),
whereas raised fasting glucose and TAGs
were observed less frequently (men: 26.9
and 63.3%; women: 26.9 and 30.6%, re-
spectively).
Under the new International Diabetes
Federation consensus statement criteria,
subjects with high waist circumference
(men: 76.8%; women: 56.1%), raised
TAGs (men: 90.9%; women: 94.3%), re-
duced HDL cholesterol (men: 52%;
women: 53.9%), raised blood pressure
(men: 38.6%; women: 63%), and raised
Table 1dPrevalence of risk factors for metabolic syndrome by sex
Risk factors for metabolic syndrome
N 272
High waist circumference*
Reduced HDL cholesterol
Raised TAGs
Elevated blood pressurex
Elevated fasting glucose||
Data are percentages, unless otherwise indicated. *Waist circumference $94 cm for men or $80 cm
for women. HDL cholesterol ,1.0 mmol/L for men or ,1.3 mmol/L for women or specic treatment for
this abnormality. TAG .1.7 mmol/L or specic treatment for this abnormality. xElevated blood pressure
.130/85 mmHg or treatment of previously diagnosed hypertension. ||Elevated fasting glucose .5.6 mmol/L
or treatment of previously diagnosed diabetes.
care.diabetesjournals.org
fasting blood glucose (men: 77.8%;
women: 80.6%) were categorized as hav-
ing metabolic syndrome. However, miss-
ing under the new criteria were 48% of
men with low HDL cholesterol and 61.4%
with raised blood pressure and 43.9%
of women with high waist circumfer-
ence and 46.1% with reduced HDL
cholesterol.
The prevalence of metabolic syn-
drome among age-groups 1824, 25
34, 3544, 4554, and .55 years was
9.9, 23.8, 39.6, 46.5, and 63.5%, respec-
tively. Table 2 shows the proportions of
men and women with risk factors for met-
abolic syndrome by age categories.
Bivariate analysis
Sex difference was observed in the asso-
ciation with socioeconomic indicators
and prevalence of metabolic syndrome,
as shown in Table 3. A test for trend re-
vealed age to be signicantly associated
with occurrence of metabolic syndrome
in both men (x2 = 56.53, df = 4, P ,
0.001) and women (x2 = 20.46, df = 4,
P , 0.001). Compared with the age-
group 1824 years, the age-group 3544
years was signicantly associated with
the occurrence of metabolic syndrome
(odds ratio [OR] 27.2 [95% CI 3.7
201.3]; P = 0.001) in men and also in
women (2.9 [1.27.2]; P = 0.018). The
association also was signicant in the age-
group 4554 years in men (25.9 [3.4
198.6]; P = 0.002) and women (5.8
[2.215.3]; P , 0.001). The highest level
of signicance was observed in the age-
group $55 years in both men (99.2
[12.6783.5]; P , 0.001) and women
(6.0 [2.216.5]; P , 0.001). Socioeco-
nomic status and occurrence of metabolic
syndrome among men and women re-
vealed different measures of associa-
tion. Socioeconomic status emerged as
Men Women
267
30.1 70.3
45.2 71.8 ,0.001
20.2 13.1
72.1 57.9
10.3 13.5
Kaduka and Associates
a signicant factor in men but not in
women. In men, middle and high wealth
quintiles were signicantly associated
with the occurrence of metabolic syn-
drome (3.0 [1.18.0], P = 0.034, and
9.4 [3.524.9], P , 0.001, respectively).
Level of education emerged as a signi-
cant factor in women but not in men.
Secondary education among women was
signicantly associated with protection
from metabolic syndrome (0.3 [0.20.8];
P = 0.012). Average monthly income
emerged as a signicant factor in men but
not in women, with an income ranging
between $125 and $375 USD being sig-
nicantly associated with the occurrence
of metabolic syndrome (7.0 [2.321.5];
P = 0.001). No signicant association
was observed between the main source
of income and the occurrence of metabolic
syndrome in both men and women.
Multivariate analysis
Binary logistic regression using the back-
ward conditional method was performed
on multiple factors to eliminate con-
founding and to examine the effect of
independent predictors on the occur-
rence of metabolic syndrome. Four fac-
tors, namely age, socioeconomic status,
level of education, and average monthly
income, found to be associated with the
occurrence of metabolic syndrome at P ,
0.1 during bivariate analysis were consid-
ered for multivariate analysis. Table 4
shows factors found to predict occurrence
of metabolic syndrome among men and
women. Occurrence of metabolic syn-
drome in men was signicantly associated
with age-groups 3544 years (adjusted
OR 39.8 [95% CI 3.4263.0]; P =
0.002), 4554 years (46.0 [4.7446.6];
P = 0.001), and $55 years (213.2
[19.52326.9]; P , 0.001). In women,
occurrence of metabolic syndrome was
signicantly associated with age-groups
4554 years (4.8 [1.415.8]; P = 0.011)
and $55 years (4.8 [1.317.8]; P = 0.018)
and not with age-group 3544 years (2.6
[0.88.3]; P = 0.097). Occurrence of met-
P abolic syndrome in men was signicantly
associated with the middle (OR 6.3 [95%
CI 1.625.1]; P = 0.009) and highest
,0.001
quintiles (adjusted OR 14.9 [95% CI
3.562.6]; P , 0.001), whereas in
0.155
women, the occurrence of metabolic syn-
,0.001
drome was signicantly associated with
0.331
the fourth wealth quintile (3.7 [1.49.8];
P = 0.009). Level of education was a sig-
nicant factor in men. Occurrence of meta-
bolic syndrome was signicantly associated
with men who had attained a university
DIABETES CARE, VOLUME 35, APRIL 2012 889
Metabolic syndrome in urban Kenya

Table 2dProportions of risk factors for metabolic syndrome among men and women by age-group

1824 years 2534 years 3544 years 4554 years $55 years
Risk factors Male Female Male Female Male Female Male Female Male Female
High waist circumference* 2 (4.2) 15 (35.7) 15 (21.1) 45 (62.5) 22 (30.1) 57 (80.3) 17 (38.6) 38 (90.5) 27 (73.0) 32 (86.5)
Reduced HDL cholesterol 24 (50.0) 23 (53.5) 26 (37.1) 58 (80.6) 37 (50.7) 54 (76.1) 14 (31.8) 28 (66.7) 21 (58.3) 27 (73.0)
Raised TAGs 1 (2.1) 2 (4.7) 6 (8.6) 4 (5.6) 23 (31.5) 10 (14.1) 13 (28.9) 9 (21.4) 13 (35.1) 9 (24.3)
Elevated blood pressurex 35 (72.9) 16 (37.2) 37 (52.9) 35 (48.6) 51 (69.9) 38 (53.5) 40 (90.9) 34 (81.0) 33 (91.7) 30 (81.1)
Elevated fasting glucose|| 0 (0.0) 1 (2.4) 5 (7.1) 3 (4.2) 5 (6.8) 12 (16.9) 7 (15.9) 10 (23.8) 11 (29.7) 9 (24.3)
Data are n (%). *Waist circumference $94 cm for men or $80 cm for women. HDL cholesterol ,1.0 mmol/L for men or ,1.3 mmol/L for women or specic
treatment for this abnormality. TAG .1.7 mmol/L or specic treatment for this abnormality. xElevated blood pressure .130/85 mmHg or treatment of previously
diagnosed hypertension. ||Elevated fasting glucose .5.6 mmol/L or treatment of previously diagnosed diabetes.

level of education (9.8 [1.186.2]; P =
0.040).
CONCLUSIONSdThe epidemiologi-
cal transition that often accompanies ur-
banization is characterized by increasing
prevalence of etiological risk factors for
metabolic syndrome, such as obesity,
diabetes, and high blood pressure. The
clustering of these risk factors confers a
greater risk of premature morbidity and
mortality (11). Our study found the over-
all prevalence of metabolic syndrome at
34.6% and was higher in women than
men (40.2 vs. 29%, respectively). Similar
patterns have been observed elsewhere in
Africa. The prevalence of metabolic syn-
drome among West African women and
men was found to be 42 vs.19%, respec-
tively; Great Tunis 31.2% (women 37.3
vs. men 23.9%, respectively); and
Seychelles between 25 and 30% (12
14). The observed sex difference in the dis-
tribution of metabolic syndrome in this
study could be attributed to differences in
socioeconomic status, attainment of higher
education, and the interplay of the two.
Education attainment has a strong effect
on health behaviors and attitudes and,
consequently, lifestyle. According to the
Kenya Demographic and Health Survey
20082009 report, a sex difference exists
in educational attainment from the age of
14 years, with more male than female
youths attending school (15). The same
report shows a direct association between
educational attainment and improvement
in wealth status for both men and women
in Kenya. In this study, the presence of
metabolic syndrome was inversely associ-
ated with attainment of higher education
in women but not in men. There was an
association between advancement in so-
cioeconomic status in both sexes, and at-
tainment of university education in men
with presence of metabolic syndrome.
Thus, there could be other factors in
addition to socioeconomic status and ed-
ucation contributing to the difference in
the associations observed that need fur-
ther investigation.
Low HDL cholesterol and high waist
circumference were more prevalent in
women as opposed to high blood pressure
in men. Studies in Brazzaville, Cameroon,
and Ghana also have observed high car-
diometabolic risk in women than men,
with central obesity and high blood
pressure as the most prevalent risk factors
(1618). Likewise, black populations in
the Caribbean, America, and Europe
have been found to have low HDL choles-
terol levels and increased waist circumfer-
ence as the common risk factors for
metabolic syndrome, and emphasis has
been on controlling central obesity and
ethnic-specic reformulation of the meta-
bolic syndrome (12,19,20). Metabolic
syndrome presents an increased risk for
type 2 diabetes and CVD. Thus, in addi-
tion to pharmacologic treatment, lifestyle
interventions, such as atherogenic diets,
weight loss, and increased physical activ-
ity, should be encouraged to address the
metabolic risk factors. There is need for
more research to establish the true burden
and causes of CVDs and diabetes across
the country. Additional qualitative re-
search will help better understand the as-
sociations observed. The inclusion of
NCDs in national surveys such as the
Kenya Demographic Health Surveys
would serve as a start.
In Kenya, chronic diseases have not
received much priority because of the
overemphasis on communicable diseases
and donor-driven agenda that prioritize
infectious diseases. Yet, ~53% of all hos-
pital admissions in Nairobi are attributed
to NCDs, and diabetes contributes
~27.3% of the total (3). It is important
for the policy makers and planners in
Kenya to be cognizant of the overlap be-
tween infectious diseases and NCDs.
In South Africa for instance, a study by
Sliwa et al. (21) observed a high prevalence
of modiable risk factors for atherosclerotic
disease and a combination of infectious
and noncommunicable forms of heart
disease. They noted that epidemiological
transition has broadened the complexity
and spectrum of heart disease in the urban
African community. In Kenya, an autopsy
study by Ogengo et al. (22) found com-
mon conditions associated with cardio-
vascular deaths, such as myocardial
infarction, cardiomyopathy, subarach-
noid hemorrhage, pulmonary thrombo-
embolism, ruptured aortic aneurysm,
hypertensive heart disease, infective peri-
carditis, and rheumatic heart disease, sug-
gesting that NCDs do overlap with
infectious conditions as causes of cardio-
vascular mortality.
Various anthropometric and bio-
chemical factors differ with different eth-
nic populations and with sex. In our
study, up to 48% of men with low HDL
cholesterol and 61.4% with raised blood
pressure and 43.9% of women with high
waist circumference and 46.1% with
reduced HDL cholesterol were missed
under the new International Diabetes
Federation consensus statement criteria.
The new criteria may fall short in this
population and result in failure to
appropriately identify individuals for
primary prevention and management
therapy (23). There is urgent need,
therefore, for prospective research data
from Africa that will guide the develop-
ment of risk assessment tools that are pop-
ulation specic and sensitive to ethnic
differences.
Identication of the social and eco-
nomic characteristics associated with oc-
currence of metabolic syndrome is essential
for the success of primary preventive mea-
sures. In women, the odds favoring meta-
bolic syndrome are said to signicantly
increase with age and in unfavorable social

890 DIABETES CARE, VOLUME 35, APRIL 2012 care.diabetesjournals.org

Kaduka and Associates

891
Table 3dAssociation of selected risk factors with metabolic syndrome among men and women in an urban population in Kenya

DIABETES CARE, VOLUME 35, APRIL 2012

Men Women
With metabolic Without metabolic With metabolic Without metabolic
syndrome (n = 79) syndrome (n = 193) syndrome (n = 107) syndrome (n = 160)
OR OR
Factors n (%) n (%) (95% CI) P n (%) n (%) (95% CI) P
Age-group (years)
1824 1 (2.1) 47 (97.9) Reference 8 (18.6) 35 (81.4) Reference
2534 10 (14.1) 61 (85.9) 7.3 (0.956.9) 0.059 24 (33.3) 48 (66.7) 2.1 (0.85.2) 0.111
3544 27 (37.0) 46 (63.0) 27.2 (3.7201.3) 0.001 29 (40.8) 42 (59.2) 2.9 (1.27.2) 0.018
4554 16 (36.4) 28 (63.6) 25.9 (3.4198.6) 0.002 24 (57.1) 18 (42.9) 5.8 (2.215.3) ,0.001
.55 25 (69.4) 11 (30.6) 99.2 (12.6783.5) ,0.001 22 (59.5) 15 (40.5) 6.0 (2.216.5) ,0.001
NR/NA/DK 0 0 0 1
Socioeconomic status
Lowest quintile 6 (11.8) 45 (88.2) Reference 35 (40.2) 52 (59.8) Reference
Second quintile 9 (21.4) 33 (78.6) 2.1 (0.76.4) 0.197 16 (38.1) 26 (61.9) 0.9 (0.42) 0.882
Middle quintile 19 (28.4) 48 (71.6) 2.9 (1.18) 0.034 29 (38.2) 47 (61.8) 0.9 (0.51.7) 0.747
Fourth quintile 8 (17.4) 38 (82.6) 1.7 (0.55.1) 0.383 21 (56.8) 16 (43.2) 2.0 (0.94.4) 0.079
Highest quintile 36 (55.4) 29 (44.6) 9.3 (3.524.9) ,0.001 6 (24.0) 19 (76.0) 0.5 (0.21.3) 0.131
Level of education
None 7 (43.8) 9 (56.3) Reference 18 (50.0) 18 (50.0) Reference
Primary 21 (21.6) 76 (78.4) 0.4 (0.11.1) 0.066 61 (45.5) 73 (54.5) 0.8 (0.41.8) 0.647
Secondary 28 (23.9) 89 (76.1) 0.4 (0.11.2) 0.097 19 (25.0) 57 (75.0) 0.3 (0.10.8) 0.012
University 23 (54.8) 19 (45.2) 1.6 (0.54.9) 0.449 4 (28.6) 10 (71.4) 0.4 (0.11.4) 0.153
NR/NA/DK 0 0 5 1
Monthly income (USD)
,40 5 (12.8) 34 (87.2) Reference 39 (50.0) 39 (50.0) Reference
4075 11 (22.0) 39 (78.0) 2.1 (0.66.7) 0.216 22 (40.0) 33 (60.0) 0.7 (0.31.3) 0.262
75125 13 (19.7) 53 (80.3) 1.8 (0.65.7) 0.295 15 (35.7) 27 (64.3) 0.6 (0.31.2) 0.164
125375 22 (48.9) 23 (51.1) 7.0 (2.321.5) 0.001 11 (52.4) 10 (47.6) 1.2 (0.43.1) 0.759
.375 24 (63.2) 14 (36.8) 12.7 (4.040.7) ,0.001 4 (33.3) 8 (66.7) 0.5 (0.11.7) 0.249
NR/NA/DK 3 31 15 41
Main source of income
Formally employed 30 (29.4) 72 (70.6) Reference 20 (40.8) 29 (59.2) Reference
None 0 (0) 4 (100) UD 0.999 3 (30.0) 7 (70.0) 0.6 (0.12.5) 0.442

care.diabetesjournals.org
Farming 2 (33.3) 4 (66.7) 1.4 (0.37.7) 0.688 4 (66.7) 2 (33.3) 4.0 (0.626.4) 0.147
Self-employed 31 (32.3) 65 (67.7) 1.2 (0.62.1) 0.639 45 (45.0) 55 (55.0) 1.2 (0.62.5) 0.542
Petty trade 3 (14.3) 18 (85.7) 0.4 (0.11.6) 0.207 15 (41.7) 21 (58.3) 1.1 (0.42.6) 0.867
Other 9 (39.1) 14 (60.9) 1.5 (0.63.9) 0.388 13 (39.4) 20 (60.6) 1.0 (0.42.4) 0.924
NR/NA/DK 3 17 9 24
DK, do not know; NA, not applicable; NR, no response; UD, undened.
Metabolic syndrome in urban Kenya
Table 4dRisk factors for the occurrence of metabolic syndrome among men and women
in an urban population in Kenya
Men
Risk factors Adjusted OR (95% CI)
Age (years)
1824 Reference
2534 9 (0.436.5)
3544 29.8 (3.4263.0)
4554 46 (4.7446.6)
.55 213.2 (19.52,326.9)
Socioeconomic status
Lowest wealth quintile Reference
Second wealth quintile 3.1 (0.813.1)
Middle wealth quintile 6.3 (1.625.1)
Fourth wealth quintile 1.9 (0.58.0)
Highest wealth quintile 14.9 (3.562.5)
Level of education
None Reference
Primary 1.7 (0.311.4)
Secondary 1.8 (0.312.3)
University 9.8 (1.186.2)
class, as described by occupation and
decreased education level. In men, meta-
bolic syndrome increases signicantly
with increasing age and socioeconomic
status (24). Our ndings were in support
of this observation. Metabolic syndrome
was prevalent among men of the highest
socioeconomic status, those with second-
ary education, those who were self-employed,
and those of high-income earnings.
Women who presented with metabolic
syndrome were mostly those from the
lowest socioeconomic status, those hav-
ing attained only primary education,
those who were self-employed, and
those of lower-income earnings. The high
incidence of metabolic syndrome ob-
served among women at lower quintiles
probably may be a result of low literacy,
limited exposure to sources of relevant in-
formation, and/or comprehension of this
information (25). Income inequality is an
independent contributor to variations
in CVD burden globally. According to
Salsberry et al. (26), women in the lowest
economic group are more likely to be at
risk when compared with women in the
highest economic group. It also is impor-
tant to note that, in certain parts of Africa,
cultural values and positive social attitudes
toward fatness among women largely con-
tributes to feminine obesity and, conse-
quently, metabolic syndrome (27). Thus,
interventions must be devised based on
the level of economic development, the
socioeconomic context of risk factor ex-
posures, and individual characteristics,
892 DIABETES CARE, VOLUME 35, APRIL 2012
Women
P Adjusted OR (95% CI) P of Southampton, who awarded the scholarship
Reference
0.230 2.4 (0.87.4) 0.139
0.002 2.6 (0.88.3) 0.097
0.001 4.8 (1.415.8) 0.011
,0.001 4.8 (1.317.8) 0.018
Reference
0.115 1.2 (0.52.8) 0.751
0.009 1.2 (0.62.6) 0.614
0.372 3.7 (1.49.8) 0.009
,0.001 0.7 (0.22.1) 0.499
0.581
0.540
0.040
such as age, sex, income, and education
level.
This study helps to create awareness
among Kenyan policy makers and plan-
ners on the health status of the Kenyan
urban population with respect to CVDs
and calls for attention to both behavioral
and biological CVD risk factors. This study
also contributes to the worldwide map-
ping of metabolic syndrome. This study
had a limitation. The cross-sectional de-
sign does not allow causal or directional
inferences. Longitudinal studies are there-
fore encouraged.
The prevalence of CVD risk factors
among urban dwellers in Kenya remains
high. This calls for action to prevent and
treat the components of the metabolic
syndrome. It is essential that the com-
munity is mindful of the important CVD
risk factors to help improve on adher-
ence to lifestyle changes and medication
among those with established disease.
The Kenyan government should encour-
age development and sustenance of health
promotion programs that will use a health
education approach focused on major
biomedical and behavioral risk factors for
development of CVD, while paying atten-
tion to both promoters and barriers to
change. The primary health care system
needs to be strengthened and integrate the
care of CVDs and management of risk
factors. Additional research is necessary to
establish the true burden of diabetes,
CVDs, and the concomitant risk factors
across the country.
AcknowledgmentsdThis study was sup-
ported in part by a scholarship from the Gregory
Taylor Memorial Fund (U.K.). The authors
thank Prof. Alan Jackson and Dr. Steven Wotton
of the Institute of Human Nutrition, University
for travel and biochemical analyses costs.
No potential conicts of interest relevant to
this article were reported.
L.U.K. researched data and wrote the man-
uscript. Y.K., J.K.B., and Z.N.B. researched
data, contributed to discussion, and reviewed
the manuscript. E.Ke. and E.Ku. contributed
to discussion and reviewed the manuscript.
M.M. contributed to discussion. L.U.K. is the
guarantor of this work and, as such, had full
access to all the data in the study and takes
responsibility for the integrity of the data and
the accuracy of the data analysis.
Parts of this study were presented in abstract
form at the 4th Africa Nutritional Epidemiol-
ogy Conference, Nairobi, Kenya, 48 October
2010.
The authors acknowledge the editorial assis-
tance of Christine Bukania, formerly of the Aga
Khan Foundation, Kenya. The authors especially
thank Dr. John Jackson and Christian Gelauf of
the Institute of Human Nutrition, University of
Southampton (U.K.), for their assistance with
the biochemical analysis, as well as Ezekiel
Ogutu and Alex Muyera of the Kenya National
Bureau of Statistics for their assistance in map-
ping the study area. The authors also thank Ruth
Kayima and Linda Atieno of the University of
Nairobi, Salma Abdi and Tobias Oliech of
KEMRI for their assistance with nutritional
assessments, and Saidi Kisiwa and Ezekiel
Mukhaye of KEMRI for their laboratory support.
The authors sincerely thank the administrative
ofcers and participants of Langata Constitu-
ency for their assistance and collaboration.
References
1. Murray CJL, Lopez AD. Global Health
Statistics: Global Burden of Disease and In-
jury Series. Boston, MA, Harvard School of
Public Health, 1996
2. Abegunde DO, Mathers CD, Adam T,
Ortegon M, Strong K. The burden and
costs of chronic diseases in low-income
and middle-income countries. Lancet 2007;
370:19291938
3. Republic of Kenya Ministry of Health.
Annual Health Sector Status Report 2005
2007. Nairobi, Kenya, Ministry of Health,
2007
4. Yonga GO, Ogola EN, Juma FD. Cardio-
vascular risk factor proles in mild to
moderate hypertensives seen at Kenyatta
National Hospital. East Afr Med J 1993;
70:693695
5. Christensen DL, Eis J, Hansen AW, et al.
Obesity and regional fat distribution in
Kenyan populations: impact of ethnicity
and urbanization. Ann Hum Biol 2008;
35:232249
care.diabetesjournals.org

6. Christensen DL, Friis H, Mwaniki DL,
et al. Prevalence of glucose intolerance
and associated risk factors in rural and
urban populations of different ethnic
groups in Kenya. Diabetes Res Clin Pr
2009;84: 303310
7. Sharma P, Mishra S. Metabolic syndrome:
early identication prevents type ii di-
abetes and cardiovascular disease. Indian J
Clin Biochem 2007;22:13
8. Kenya National Bureau of Statistics, Re-
public of Kenya. The 2009 Kenya Pop-
ulation and Housing Census, Volume 1B:
Population Distribution by Political Units,
Nairobi, Kenya, Kenya National Bureau of
Statistics, 2010
9. McBurney P. On transferring statistical
techniques across cultures: the Kish Grid.
Curr Anthropol 1988;29:323325
10. Alberti KG, Eckel RH, Grundy SM, et al.;
International Diabetes Federation Task
Force on Epidemiology and Prevention;
National Heart, Lung, and Blood Institute;
American Heart Association; World Heart
Federation; International Atherosclerosis
Society; International Association for the
Study of Obesity. Harmonizing the me-
tabolic syndrome: a joint interim statement
of the International Diabetes Federation
Task Force on Epidemiology and Pre-
vention; National Heart, Lung, and Blood
Institute; American Heart Association;
World Heart Federation; International
Atherosclerosis Society; and International
Association for the Study of Obesity. Cir-
culation 2009;120:16401645
11. Lakka HM, Laaksonen DE, Lakka TA,
et al. The metabolic syndrome and total
and cardiovascular disease mortality in
middle-aged men. JAMA 2002;288:2709
2716
12. Sumner AE, Zhou J, Doumatey A, et al.
Low HDL-cholesterol with normal tri-
glyceride levels is the most common lipid
pattern in West Africans and African
Americans with metabolic syndrome:
care.diabetesjournals.org
implications for cardiovascular disease
prevention. CVD Prev Contr 2010;5:
7580
13. Allal-Elasmi M, Haj Taieb S, Hsairi M,
et al. The metabolic syndrome: preva-
lence, main characteristics and association
with socio-economic status in adults liv-
ing in Great Tunis. Diabetes Metab 2010;
36:204208
14. Kelliny C, William J, Riesen W, Paccaud F,
Bovet P. Metabolic syndrome according
to different denitions in a rapidly de-
veloping country of the African region.
Cardiovasc Diabetol 2008;7:27
15. Kenya National Bureau of Statistics
(KNBS) and ICF Macro. Kenya Demo-
graphic and Health Survey Report 2008
2009. Calverton, Maryland, KNBS and
ICF Macro, 2010
16. Fezeu L, Balkau B, Kengne AP, Sobngwi E,
Mbanya JC. Metabolic syndrome in a sub-
Saharan African setting: central obesity
may be the key determinant. Atheroscle-
rosis 2007;193:7076
17. Agyemang C. Rural and urban differences
in blood pressure and hypertension in
Ghana, West Africa. Public Health 2006;
120:525533
18. Gombet T, Longo-Mbenza B, Ellenga-Mbolla
B, Ikama MS, Kimbally-Kaky G, Nkoua JL.
Relationship between coronary heart dis-
ease, metabolic syndrome, energy expen-
diture, body composition, kidney function
and low-grade inammation among bank
African employees in Brazzaville. Diabetes
Metab Syndr Clin Res Rev 2010;4:197
203
19. Ezenwaka CE, Nwagbara E, Seales D, et al.
A comparative study of the prevalence of
the metabolic syndrome and its compo-
nents in type 2 diabetic patients in two
Caribbean islands using the new Inter-
national Diabetes Federation denition.
Arch Physiol Biochem 2007;113:202210
20. Meis SB, Schuster D, Gaillard T, Osei K.
Metabolic syndrome in nondiabetic, obese,
Kaduka and Associates
rst-degree relatives of African American
patients with type 2 diabetes: African
American triglycerides-HDL-C and in-
sulin resistance paradox. Ethn Dis 2006;
16:830836
21. Sliwa K, Wilkinson D, Hansen C, et al.
Spectrum of heart disease and risk fac-
tors in a black urban population in
South Africa (the Heart of Soweto
Study): a cohort study. Lancet 2008;371:
915922
22. Ogengo JA, Gatonga P, Olabu BO. Car-
diovascular causes of death in an east Af-
rican country: an autopsy study. Cardiol J
2011;18:6772
23. Jennings CL, Lambert EV, Collins M,
Levitt NS, Goedecke JH. The atypical
presentation of the metabolic syndrome
components in black African women: the
relationship with insulin resistance and
the inuence of regional adipose tissue
distribution. Metabolism 2009;58:149
157
24. Sodjinou R, Agueh V, Fayomi B, Delisle H.
Obesity and cardio-metabolic risk factors
in urban adults of Benin: relationship with
socio-economic status, urbanisation, and
lifestyle patterns. BMC Public Health
2008;8:84
25. Scuteri A, Vuga M, Najjar SS, et al. Edu-
cation eclipses ethnicity in predicting the
development of the metabolic syndrome
in different ethnic groups in midlife: the
Study of Womens Health Across the Na-
tion (SWAN). Diabet Med 2008;25:1390
1399
26. Salsberry PJ, Corwin E, Reagan PB. A
complex web of risks for metabolic
syndrome: race/ethnicity, economics,
and gender. Am J Prev Med 2007;33:
114120
27. Holdsworth M, Gartner A, Landais E,
Maire B, Delpeuch F. Perceptions of
healthy and desirable body size in
urban Senegalese women. Int J Obes
2004;28:15611568
DIABETES CARE, VOLUME 35, APRIL 2012 893